Re: [exim] Client SMTP keepalive methods?

Αρχική Σελίδα
Delete this message
Reply to this message
Συντάκτης: Alan J. Flavell
Ημερομηνία:  
Προς: Exim users list
Αντικείμενο: Re: [exim] Client SMTP keepalive methods?
On Tue, 6 Jun 2006, Chris Blaise wrote:

[unattributed quote - author was in fact Tony Finch:]

> > Note that the SMTP standard requires clients to wait at least
> > 10 minutes for the post-DATA response. If Outlook only waits
> > 1 minute then it is not compliant with the specification.
> > What a surprise!
>
>     Well, no.  RFC 2821 says SHOULD in this matter, not MUST.


<quote>
  SHOULD   This word, or the adjective "RECOMMENDED", mean that
      there may exist valid reasons in particular circumstances to
      ignore a particular item, but the full implications must be
      understood and carefully weighed before choosing a different
      course.
</>


So: the users of such software are deemed to have considered the full
implications, and decided that they accepted the consequences of their
choice.

Who are we to argue? As long as *we* respond as soon as we can, and
in any case within the limits set by the specification, the
consequences are *theirs* - which they, by definition, made with "the
full implications understood and carefully weighed".

> I'm not saying I like it, but it's the reality I have to deal
> with...


The reality is that email is not guaranteed - and its reliability is
getting worse and worse, due to defences (some good, some bad) that
have to be erected against the torrent of abuse. It does *not* get
better by having to make allowances for violations of the published
interworking specifications; and, every time that we accommodate to
one of those violations, by otherwise bona-fide senders, we *delay*
the correction of that part of the problem. That approach is
demonstrably not working. Unless we stand together and insist on
specification compliance, it's going to just get worse, and worse.

Senders who give up prematurely in the face of a modest delay
("modest" in terms of the published specification, I mean) are usually
spammers. I don't see any reason not to treat anyone behaving like
that as suspects.

Sorry, but that's the way I see it.